1. Field of the Invention
The present invention relates to an output method for improving image quality. More particularly, the present invention relates to an output method for improving video image quality for image compression.
2. Description of Related Art
In general, the amount of digital video data is very large. Therefore, in order to store the video data within a smaller memory space and maintain high speed video data transmission, the video data has to be compressed. Conventionally, the data compression is a process for removing the unnecessary information of video data to reduce the amount of data. For example, if previous and next frames are similar, the previous frame may be retained and the portion, existing in both frames, of the next frame may be removed (i.e., only the portions of the next frame, different from the previous frame are retained). Accordingly, the amount of digital video data may be significantly reduced.
Conventionally, in the motion picture experts group (MPEG) specification, the decoding of video image may generally produce three types of frames, namely, intra-coding frame (I frame), prediction frame (P frame) and bidirectional frame (B frame) respectively. In the decoding process of high definition digital television or high definition digital video disc (e.g., digital versatile disc, DVD), the memory capacity for storing frame data of I frame, P frame and B frame is very large. If the memory capacity is limited, the frame data stored in the memory has to be compressed appropriately to release a couple of memory space for other decoding process. Thereafter, if the frame data of the I frame and the P frame has to be compressed simultaneously, the image quality of the B frame decoded according to the compressed I frame and P frame are reduced. Therefore, the image quality of the picture group is seriously deteriorated. According to the reason described above, only the frame data of the B frame is compressed in the conventional process.
FIG. 1 is an output procedure of a conventional video data. Referring to FIG. 1, in the conventional output procedure of a video data, the decoded sequence 10 sequentially includes I frame 110a, P frame 120a, B frames 130a and 130b, P frame 120b, B frames 130c and 130d, P frame 120c, B frames 130e and 130f. It is noted that, the I frame 110a, P frames 120a, 120b and 120c is coincide with the first standard, and the B frames 130a, 130b, 130c, 130d, 130e and 130f coincide with the second standard. The data amount of the first standard is generally larger than that of the second standard. In addition, the second standard may be, for example, the conventional phase alternation by line (PAL) standard or the national television system committee (NTSC) standard.
The frame data of the decoding sequence 10 has to be rearranged to obtain the frame data of a display sequence 20, wherein the display sequence 20 includes I frame 110a, B frames 132a and 132b, P frame 120a, B frames 132c and 132d, P frame 120b, B frames 132e and 132f and P frame 120c sequentially. It is noted that, the B frames 130a, 130b, 130c, 130d, 130e and 130f with the second standard have to be decompressed via a decompression process to obtain the B frames 132a, 132b, 132c, 132d, 132e and 132f with the first standard performed in the display sequence 20.
In the display sequence 20, the B frames 132a, 132b, 132c, 132d, 132e and 132f has smaller S/N ratio (signal to noise ratio) compared to the I frame 110a and the P frames 120a, 120b and 120c. In other words, the B frames 132a, 132b, 132c, 132d, 132e and 132f are more blurred than the I frame 110a and the P frames 120a, 120b and 120c. Therefore, if only the frame data of the B frame is compressed, the image quality is not uniform and may look like sometimes clear and sometimes blurred.